A robust matrix root-clustering technique for vehicle dynamics : the output feedback case

This paper deals with the active control in vehicles to improve safety and stability. A 4 WS vehicle model based on the roll, yaw and lateral dynamics is used. In design, the vehicle dynamics are represented by a LPV (linear parameter varying) model subject to external disturbances. According to this, the corresponding state-space model is affected by both polytopic and norm-bounded uncertainties. It is aimed to compute K, associated with a robust static output feedback, so as to place the set of all eigenvalues of the uncertain closed-loop state matrix A_cl in a clustering region, denoted by D_C, while minimizing the effect of the lateral disturbances on the system and ensuring some transient performances. This paper proposes a heuristic resolution to compute K. It is based on genetic algorithms (nonhere detailed) and on the resolution of LMI. In order to highlight the performance of the proposed control algorithm, numerical simulations are performed.